Akademik Veri Yönetim Sistemi
3rd International Conference on Computer Science and Engineering (UBMK), Sarajevo, Bosna-Hersek, 20 - 23 Eylül 2018, ss.239-243
With the development of technology, there has been an increase in production and the number of accidents has increased. Progressive technology, the development of the industry and the lack of protective precaution, and the responsibility for the uneducated employees are the main causes of work accidents. In this study, the types of injuries in work accidents and the effects of the injuries on the body are analyzed via Bayesian Networks (BNs). The BNs reflect the conditional dependency relations between variables and, the fact that they are not dependent on a single independent variable. BNs are constructed on a dataset from an international construction company. The accuracy rate and other performance measures of the constructed Bayesian network are analyzed and the effectiveness of the constructed model is analyzed. According to the experimental results, it's explicit that some cases of job accidents can be predicted beforehand with high accuracies by using machine learning techniques.